the initial wave of state polls showed romney doing pretty well in the battleground states. in the rasmussen polls over the last two days, obama is doing better. that's consistent with the trend i talked about earlier. the question is whether obama gets an outsized bounce in states like ohio, nevada and iowa where he was doing better than nationally. if you look at all of the national polls conducted since the first debate, what you see is about a three point gain for romney. if that holds and if that uniformly filtering down to the battleground states, romney would still trail in iowa, ohio and alabama and obama would have a slight advantage ♪ electoral college. >> continuing our theme of what really matters, let's talk about the polls. obviously, polls don't vote. but do they sway the electorate? do they see one person as leading or gaining and some people decide to change their vote? is there some science on do polls actually shape the election itself? >> i don't think there's very much evidence that polls shift the election. if they did, i think we would see, you know,